1 / 23

Extensive-form games

Extensive-form games. Extensive-form games with perfect information. Players do not move simultaneously When moving, each player is aware of all the previous moves ( perfect information ) A (pure) strategy for player i is a mapping from player i’s nodes to actions (when at X do s1).

shamus
Download Presentation

Extensive-form games

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Extensive-form games

  2. Extensive-form games with perfect information • Players do not move simultaneously • When moving, each player is aware of all the previous moves (perfect information) • A (pure) strategy for player i is a mapping from player i’s nodes to actions (when at X do s1) Player 1 Player 2 Player 2 Player 1 2, 4 5, 3 3, 2 Leaves of the tree show player 1’s utility first, then player 2’s utility 1, 0 0, 5

  3. Backward induction • When we know what will happen at each of a node’s children, we can decide the best action for the player who is moving at that node • Substitute the result of the computation at the parent node. Player 1 Player 2 Player 2 Player 1 1,0 3, 2 2, 4 5, 3 1, 0 0, 5

  4. What happens with backwards induction here?

  5. A limitation of backward induction • Tree may be too long to evaluate. • If there are ties, then how they are broken affects what happens higher up in the tree • Colors show multiple possible strategies. • Multiple equilibria… Player 1 .12345 .87655 Player 2 Player 2 1/2 1/2 3, 2 2, 3 4, 1 0, 1

  6. Conversion from extensive to normal form player 1 is row player LR = Left if 1 moves Left, Right if 1 moves Right; etc. Player 1 LL LR RL RR Left Right Left Player 2 Player 2 Right L L R R • Nash equilibria of this normal-form game include (Right, LL), (Right, RL), (Left, RR) + infinitely many mixed-strategy equilibria • In general, normal form can have exponentially many strategies 3, 2 2, 3 4, 1 0, 1 Underscores mean, given you know what the other person will do, what will you do?

  7. Conversion from extensive to normal form player 1 is row player for player 2 LR = Left if 1 moves Left, Right if 1 moves Right; etc. Player 1 LL LR RL RR Left Player 2 Player 2 Right One concern, why is column player playing LL? It doesn’t appear to be a smart strategy. Will player 1 believe him or is he claiming that strategy to get an advantage? 3, 2 2, 3 4, 1 0, 1 Underscores mean, given you know what the other person will do, what will you do?

  8. Theorem 5.1.3 (SLB) Every finite perfect-information game in extensive form has a pure-strategy equilibrium. Why?

  9. Converting another game to normal form Player 1 Do at your seats… Player 2 Player 2 Player 1 2, 4 5, 3 3, 2 1, 0 0, 1

  10. Converting another game to normal form LL LR RL RR Player 1 LL LR RL Player 2 Player 2 RR • Pure-strategy Nash equilibria of this game are (LL, LR), (LR, LR), (RL, LL), (RR, LL) • But the only backward induction solution is (RL, LL) • Why? Player 1 2, 4 5, 3 3, 2 1, 0 0, 1

  11. Subgame perfect equilibrium • Each node in a (perfect-information) game tree, together with the remainder of the game after that node is reached, is called a subgame • A strategy profile is a subgame perfect equilibrium if it is an equilibrium for every subgame LL LR RL RR LL Player 1 LR RL RR Player 2 Player 2 *L *R ** *L *L *R *R Player 1 • (RR, LL) and (LR, LR) are not subgame perfect equilibria because (*R, **) is not an equilibrium as player 1 shouldn’t play R at second node. • (LL, LR) is not subgame perfect because (*L, *R) is not an equilibrium • *R for player 2 is not a credible threat 2, 4 5, 3 3, 2 1, 0 0, 1

  12. Imperfect information • Dotted lines indicate that a player cannot distinguish between two (or more) states • A set of states that are connected by dotted lines is called an information set. We consider moves from anyplace in the set as a unit • Reflected in the normal-form representation Player 1 At seats, what is normal form? Show both extensive form and normal form as want to show comparisons between methods Player 2 Player 2 0, 0 -1, 1 1, -1 -5, -5 • Any normal-form game can be transformed into an imperfect-information extensive-form game this way

  13. Imperfect information • Dotted lines indicate that a player cannot distinguish between two (or more) states • A set of states that are connected by dotted lines is called an information set • Reflected in the normal-form representation Player 1 L R L R Player 2 Player 2 Note at each node of player 2, you must make the same choice. Why? 0, 0 -1, 1 1, -1 -5, -5 • Any normal-form game can be transformed into an imperfect-information extensive-form game this way

  14. A poker-like gamePractice turning the game tree into the normal form. It is a bit tricky “nature” 2/3 1/3 1 gets King 1 gets Jack cc cf fc ff player 1 player 1 1/3 bb bet stay bet stay bs 2/3 player 2 player 2 sb call fold call fold call fold call fold ss 2 1 1 1 -2 1 -1 1 Remove dominated strategies Do mixed strategy on the others. Reward for player 1 is shown Zero sum game

  15. Computing mixed NE Column player selects q so that row player has no preference between row1 or row 2. row 1 gets 0q + 1(1-q) row 2 gets .5q + 0(1-q) to have no preference, set them equal q = 2/3 Solving for p is similar. Notice that in mixed strategies, the whole strategy is given a probability. In behavioral strategies, each node can have a different mix.

  16. Subgame perfection and imperfect information • How should we extend the notion of subgame perfection to games of imperfect information? Player 1 • We cannot expect Player 2 to play Right after Player 1 plays Left, and Left after Player 1 plays Right, because of the information set, player 2 doesn’t know which state he is in. Player 2 Player 2 1, -1 -1, 1 -1, 1 1, -1 Let us say that a subtree is a subgame only if there are no information sets that connect the subtree to parts outside the subtree

  17. Subgame perfection and imperfect information… Player 1 Player 2 Player 2 Player 2 4, 1 0, 0 5, 1 1, 0 3, 2 2, 3 • One of the Nash equilibria is: (R, RR) • Also subgame perfect (the only subgames are the whole game, and the subgame after Player 1 moves Right) • But it is not reasonable to believe that Player 2 will move Right after Player 1 moves Left/Middle (not a credible threat) • There exist more sophisticated refinements of Nash equilibrium that rule out such behavior

  18. Computing equilibria in the extensive form • Can just use normal-form representation • Misses issues of subgame perfection, etc. • Another problem: there are exponentially many pure strategies, so normal form is exponentially larger • Even given polynomial-time algorithms for normal form, time would still be exponential in the size of the extensive form as you have to build the table first. • There are other techniques that reason directly over the extensive form and scale much better • E.g., using the sequence form of the game

  19. Commitment • Consider the following (normal-form) game: • How should this game be played? • Now suppose the game is played as follows: • Player 1 commits to playing one of the rows, • Player 2 observes the commitment and then chooses a column • What is the optimal strategy for player 1? • What if 1 can commit to a mixed strategy?

  20. Commitment as an extensive-form game • For the case of committing to a pure strategy: Player 1 Up Down Player 2 Player 2 Left Right Left Right 2, 1 4, 0 1, 0 3, 1

  21. Commitment as an extensive-form game • For the case of committing to a mixed strategy: Player 1 (1,0) (=Up) (0,1) (=Down) (.5,.5) … … Player 2 Left Right Left Right Left Right 3, 1 2, 1 4, 0 1.5, .5 3.5, .5 1, 0 • Infinite-size game; computationally impractical to reason with the extensive form here

  22. Solving for the optimal mixed strategy to commit to[Conitzer & Sandholm 2006; see also: von Stengel & Zamir 2004, Letchford, Conitzer, Munagala 2009] • For every column t separately, we will solve separately for the best mixed row strategy (defined by ps) that induces player 2 to play t • maximize Σs ps u1(s, t) • subject to • for any t’, Σs ps u2(s, t) ≥Σs ps u2(s, t’) • Σs ps = 1 • (May be infeasible, e.g., if t is strictly dominated) • Pick the t that is best for player 1

  23. Visualization (0,1,0) = M C Shows what happens to utility as probability of each choice changes. A 3D version of the best response graph. R L (1,0,0) = U (0,0,1) = D

More Related